Interactive Neural Networks

Interactive Neural Networks

#explainable AI#education#machine learning#web dev

How to make machine learning more intuitive for beginners?

Abstract

AI feels like a black box because we teach symbols and terminology before developing intuition.

In this project, I turn neural networks into something you can play with and understand through interaction.

Interactive Neural Networks is an educational exploration which aims to enhance public’s understanding of neural networks by developing an interactive online guide that explains their inner workings. Through the use of text explanations, real-time simulations, and interactive graphics, users can actively engage with the learning process by modifying parameters and observing the corresponding effects.

This approach fosters a more intuitive and hands-on understanding of complex concepts, contrasted with traditional, passive learning methods. This project contributes to the principles of Explainable AI (XAI) by promoting transparency in AI decision-making process, building trust and encouraging responsible use of this technology.

Toolkit

React.js, custom machine learning model, Arduino, electronics

Exhibitions

Art and Design Education: FutureLab, 2025

West Bund Art Center, Shanghai, China

NYU Global Show & Tell, 2025

NYU Shanghai, Shanghai, China

Conference

COSA (Center for Open Source Arts) x NYU Machine Learning for Creative Coding Conference, New York, Mar 2025 Slides

Links

Live Demo: interactivenn.netlify.app

Github Repo

Demo

Explorable Explanation of a single perceptron:

Neuron Demo

Explorable Explanation of multi-layer perceptron:

MLP Demo

Video Demo:

Design

ImageImage
ImageImage

Exhibitions

ImageImage
ImageImage

Trailer Video for the Physical Installation (For Children):

Presentation at COSA x NYU Machine Learning for Creative Coding Conference

ImageImage